One row per completed germination test, with actual number sown/tested and number germinated — the site never assumes a sample size for percentage-only data. Storage age in days / months / years or from dates (exact, never rounded). Substrate + germination treatment are recommended when available: they keep tests run under different conditions from being pooled into one trajectory. Use consistent labels within a file — differently named protocols are treated as distinct (no synonym guessing). Drying treatment and test IDs are optional.
info — 2 row(s) used legacy storage="Initial" — mapped to is_initial=TRUE.
11 completed tests parsed · columns recognized: taxon, accession, isInitial, storage, substrate, treatment, ageYears, tested, germinated. Everything runs in your browser.
The run will: group your tests into formal trajectories (taxon × accession × storage × germination protocol — never pooled across protocols) → anchor each with its matched same-protocol Initial where one exists → fit binomial-probit survival on the actual counts → classify every fit (Formal survival / Exploratory / Do not interpret / No formal fit) with early-rise, plateau, and separation QC → evaluate the fitted curve against the accession's RCI P70 → detect compatible 5 °C / −18 °C pairs and fit the shared-Ki temperature comparison with evidence tiers.